4/29/2020

Motivation

  • Question

    • Is the spread of coronavirus influenced by the climate? Are warmer and sunnier places less likely to be influenced by the virus?
  • Method

    • Use latitude as a proxy for climate

    • Explore the relationship between latitude and reported cases using simple regression models

Confirmed Cases by April 20, 2020

Data Source: CSSE at Johns Hopkins University
time_series_covid19_confirmed_US.csv, time_series_covid19_confirmed_global_iso3_regions.csv

Simple Linear Model: confirmed cases = a + b * latitude

Estimate Std. Error t P
(Intercept) 2675.01 2099.04 1.27 0.20346
Latitude 216.02 63.16 3.42 0.00071

Simple Log-linear Model: log(confirmed cases) = a + b * latitude

Estimate Std. Error t P
(Intercept) 5.36 0.19 27.84 4.8e-87
Latitude 0.04 0.01 6.67 1.1e-10

Conclusion

The regression result shows the number of confirmed coronavirus cases is significantly influenced by the latitude. A one-unit increase in latitude multiplies the expected value of the number of confirmed cases by 1.04(e^0.04). The effect is significant, with a p value much less than 0.05, though it’s a minor influence in terms of the magnitude. In addition, the model is too simple to well explain the effect of climate on the spread of coronavirus cases, in next steps, more variables should be considered and included into the model.